Next generation massive access techniques for IoT systems in 6G

In this thesis document a simplified design of the sparse Interleave Division Multiple-Access (IDMA) scheme is introduced. Sparse IDMA is an advanced random access protocol for massive connectivity in the Internet of Things (IoT) context. This model was recently proposed as a new communication paradigm for the unsourced and uncoordinated Gaussian multiple access (GMAC) problem. The main feature of this encoding/decoding scheme is the joint Belief Propagation (BP) algorithm which provides simultaneous decoding of all concurrent transmissions. The key idea is to implement a low complexity version of the work proposed by Chamberland et al. in order to get as close as possible to the results of sparse IDMA while reducing the computational complexity of the scheme. Indeed, instead of the heavy joint BP, this scheme applies a simpler single-user serial Low-Density Parity-Check (LDPC) decoding algorithm through a fixed (128, 64)-LDPC code for each number of active users. Moreover, this approach should eliminate the need for complex optimization procedures. Therefore, each user’s message is divided into two parts: the first one is transmitted using Compressed Sensing (CS) to identify the preambles of the users, while the second uses LDPC transmission to recovery the data. However, despite being easy to implement, the single-user LDPC code used is particularly inefficient from a performance point of view as we will see better later. Therefore, an in-depth study will be conducted on the performance of the CS solver in order to optimize as much as possible the multi-user detection (MUD), avoiding unnecessarily stressing the data recovery part. Finally, the scheme is integrated with Successive Interference Cancellation (SIC) and Maximal Ratio Combining (MRC) for further performance improvements. This approach offers good performance with low computational complexity, as demonstrated by simulations compared to existing methods.